Becoming the 'Perf Person' in C++?

I have about 1.5 years of experience in C++ (embedded / low-level). In my team, nobody really has a strong process for performance optimization (runtime, memory, throughput, cache behavior, etc.).

I think if I build this skill, it could make me stand out. Where should I start? Which resources (books, blogs, talks, codebases) actually teach real-world performance work — including profiling, measuring, and writing cache-aware code?

Thanks.